All proofs
Project-declaredLean 4.32.0 · mathlib@81a5d257c8e4

Exceptional set carleson

exceptional_set_carleson

Plain-language statement

Let ff be 2π2\pi-periodic and belong to Lq((0,2π])L^q((0,2\pi]) for some q>1q>1. Given thresholds δ,ε>0\delta,\varepsilon>0, there is an index N0N_0 such that the set where the tail error supN>N0f(x)SNf(x)\sup_{N>N_0}\lVert f(x)-S_Nf(x)\rVert exceeds δ\delta has measure at most ε\varepsilon.

Exact Lean statement

theorem exceptional_set_carleson {f : ℝ → ℂ} (periodic_f : f.Periodic (2 * π))
  {q : ℝ≥0∞} (hq : 1 < q) (hf : MemLp f q (volume.restrict (Set.Ioc 0 (2 * π))))
  {δ ε : NNReal} (δpos : 0 < δ) (εpos : 0 < ε) :
    ∃ N₀, distribution (fun x ↦ ⨆ N > N₀, ‖f x - S_ N f x‖ₑ) δ (volume.restrict (Set.Ioc 0 (2 * π))) ≤ ε

Formal artifact

Lean source

Canonical source
Full Lean sourceLean 4
theorem exceptional_set_carleson {f :   ℂ} (periodic_f : f.Periodic (2 * π))  {q : 0∞} (hq : 1 < q) (hf : MemLp f q (volume.restrict (Set.Ioc 0 (2 * π))))  {δ ε : NNReal} (δpos : 0 < δ) (εpos : 0 < ε) :     N₀, distribution (fun x  ⨆ N > N₀, ‖f x - S_ N f x‖ₑ) δ (volume.restrict (Set.Ioc 0 (2 * π)))  ε := by  set p := (min (3 / 2) q).toNNReal with p_def  have : min (3 / 2) q := by    apply ne_of_lt    rw [min_lt_iff]    left    refine div_lt_top (by simp) (by simp)  have hp : p  Set.Ioo 1 2 := by    rw [p_def]    simp only [Set.mem_Ioo]    constructor    · rw [ coe_lt_iff_lt_toNNReal this, lt_min_iff]      simp only [ENNReal.coe_one]      symm      use hq      exact (ENNReal.lt_div_iff_mul_lt (by simp) (by simp)).mpr (by simp; norm_num)    · apply toNNReal_lt_of_lt_coe      simp only [coe_ofNat, inf_lt_iff]      left      refine div_lt_of_lt_mul ?_      norm_num  have hf : MemLp f p (volume.restrict (Set.Ioc 0 (2 * π))) := by    apply MemLp.mono_exponent hf    rw [p_def, coe_toNNReal this]    exact min_le_right _ _  have meas_f : AEStronglyMeasurable f :=    periodic_f.aestronglyMeasurable (t := 0) (by simp [hf.1])  have one_lt_p : (1 : ENNReal) < p := by simp [hp.1]  have δ2pos : 0 < δ / 2 := by positivity  have δ4pos : 0 < δ / 4 := by positivity  have ε2pos : 0 < ε / 2 := by positivity  have ε4pos : 0 < ε / 4 := by positivity  /- Approximate f by a smooth f₀. -/  obtain f₀, contDiff_f₀, periodic_f₀, hf₀ :=    close_smooth_approx_periodic_Lp one_lt_p.le (by simp) hf      (lt_min (C_distribution_le_of_eLpNorm_le_pos (p := p) δ4pos ε2pos)        (C_control_approximation_effect_pos (p := p) δ2pos ε2pos))  obtain N₀, hN₀ := fourierConv_ofTwiceDifferentiable periodic_f₀    ((contDiff_infty.mp (contDiff_f₀)) 2) δ4pos  /- This is a classical "epsilon third" argument. -/  use N₀  have : ᶠx in (ae (volume.restrict (Set.Ioc 0 (2 * π)))), ⨆ N > N₀, ‖f x - S_ N f x‖ₑ       ‖f x - f₀ x‖ₑ + (⨆ N > N₀, ‖f₀ x - S_ N f₀ x‖ₑ) + ⨆ N, ‖S_ N (f₀ - f) x‖ₑ := by    rw [ae_restrict_iff' measurableSet_Ioc]    filter_upwards with x hx    apply iSup_le    intro N    apply iSup_le    intro hN    calc ‖f x - S_ N f x‖ₑ      _ = ‖(f x - f₀ x) + (f₀ x - S_ N f₀ x) + (S_ N f₀ x - S_ N f x)‖ₑ := by ring_nf      _  ‖(f x - f₀ x) + (f₀ x - S_ N f₀ x)‖ₑ + ‖S_ N f₀ x - S_ N f x‖ₑ := enorm_add_le ..      _  ‖f x - f₀ x‖ₑ + ‖f₀ x - S_ N f₀ x‖ₑ + ‖S_ N f₀ x - S_ N f x‖ₑ :=        add_le_add_left (enorm_add_le ..) _      _  ‖f x - f₀ x‖ₑ + (⨆ N > N₀, ‖f₀ x - S_ N f₀ x‖ₑ) + ⨆ N, ‖S_ N (f₀ - f) x‖ₑ := by        gcongr        · exact le_iSup₂_of_le N hN (le_of_eq rfl)        · apply le_iSup_of_le N          rw [partialFourierSum_sub (contDiff_f₀.continuous.intervalIntegrable _ _)]          · rfl          rw [intervalIntegrable_iff_integrableOn_Ioc_of_le two_pi_pos.le]          apply hf.integrable (by simp [hp.1.le])  calc _    _  distribution        (fun x  ‖f x - f₀ x‖ₑ + (⨆ N > N₀, ‖f₀ x - S_ N f₀ x‖ₑ) + ⨆ N, ‖S_ N (f₀ - f) x‖ₑ)        ((δ / 4) +/ 4) +/ 2)) (volume.restrict (Set.Ioc 0 (2 * π))) := by      apply distribution_mono      · filter_upwards [this] with x hx        simp [hx]      · norm_cast        ring_nf        rfl    _  distribution (fun x  ‖f x - f₀ x‖ₑ) (δ / 4) (volume.restrict (Set.Ioc 0 (2 * π)))        + distribution (fun x  ⨆ N > N₀, ‖f₀ x - S_ N f₀ x‖ₑ) (δ / 4) (volume.restrict (Set.Ioc 0 (2 * π)))        + distribution (fun x  ⨆ N, ‖S_ N (f₀ - f) x‖ₑ) (δ / 2) (volume.restrict (Set.Ioc 0 (2 * π))) := by      apply distribution_add_le.trans      gcongr      exact distribution_add_le    _  ε / 2 + 0 + ε / 2 := by      gcongr      · norm_cast        apply distribution_le_of_eLpNorm_le (p := p) (by positivity) (by positivity [hp.1])          (meas_f.sub contDiff_f₀.continuous.aestronglyMeasurable).enorm.aestronglyMeasurable.restrict        simp only [eLpNorm_enorm]        apply hf₀.trans        simp      · simp only [gt_iff_lt, nonpos_iff_eq_zero]        rw [distribution_eq_zero_iff]        apply essSup_le_of_ae_le _        rw [EventuallyLE, ae_restrict_iff' measurableSet_Ioc]        filter_upwards with x hx        simp only [enorm_eq_self, iSup_le_iff]        intro N hN        rw [enorm_eq_nnnorm,  nnnorm_norm]        norm_cast        calc _          _  ‖((δ / 4) : )‖₊ := by            apply nnnorm_le_nnnorm (by simp)            exact hN₀ N hN x (Set.Ioc_subset_Icc_self hx)          _ = δ / 4 := by simp      · norm_cast        apply control_approximation_effect (ε := ε / 2) δ2pos          (contDiff_f₀.continuous.aestronglyMeasurable.sub meas_f) (periodic_f₀.sub periodic_f) hp        rw [eLpNorm_sub_comm]        apply hf₀.trans        simp    _ = ε := by simp
Project
Carleson formalization
License
Apache-2.0
Commit
74ef907d6bdb
Source
Carleson/Classical/CarlesonHunt.lean:23-132

Reuse this declaration

Bring the exact result into your workflow

The import identifies the source module. Your project still needs the pinned package dependency shown on this page.

What this badge means

This completion status comes from the project or community source. It has not yet been represented here as an independent rebuild and axiom audit.

Continue in this project

Related declarations

Project-declaredLean 4.32.0

Ae tendsto zero of distribution le

ae_tendsto_zero_of_distribution_le

Plain-language statement

Suppose that, for every error threshold δ>0\delta>0 and every measure tolerance ε>0\varepsilon>0, one can choose N0N_0 so that the set where supN>N0f(x)FN(x)\sup_{N>N_0}\lVert f(x)-F_N(x)\rVert exceeds δ\delta has measure at most ε\varepsilon. Then FN(x)F_N(x) converges to f(x)f(x) for almost every xx.

harmonic analysisFourier analysismeasure theory

Source project: Carleson formalization

Person-level attribution pending.

View proof record
Project-declaredLean 4.32.0

Antichain operator

antichain_operator

Plain-language statement

For an antichain A\mathfrak{A} of pairwise incomparable tiles, and measurable functions ff and gg bounded by the indicators of FF and GG, the pairing of gg with the Carleson sum over A\mathfrak{A} is controlled by the L2L^2 norms of ff and gg and by positive powers of the two tile-density parameters. Concretely, the bound is

C(a,q)dens1(A)(q1)/(8a4)dens2(A)1/q1/2f2g2.C(a,q)\,\mathrm{dens}_1(\mathfrak{A})^{(q-1)/(8a^4)}\,\mathrm{dens}_2(\mathfrak{A})^{1/q-1/2}\,\lVert f\rVert_2\lVert g\rVert_2.

harmonic analysisFourier analysismeasure theory

Source project: Carleson formalization

Person-level attribution pending.

View proof record